BSD Checksum Calculator

How it is calculated

The tool reads the text as bytes (UTF-8 for anything beyond plain ASCII) and keeps a 16-bit running value that starts at 0. For every byte it does two things: rotate the value right by one bit, so the lowest bit moves to the top, s = (s >> 1) + ((s & 1) << 15), and then add the byte and cut the result to 16 bits, s = (s + byte) & 0xFFFF. The final value is shown in decimal and in hexadecimal, together with the byte count. It matches the traditional BSD algorithm of the Unix sum command for the strings tested. Every character counts, including spaces and line breaks, and accented letters or emoji are several bytes. Empty text gives no result. It is a quick legacy check, not protection against tampering.

Common mistakes

• Adding or dropping a space or line break. A trailing newline is one more byte and changes the checksum completely: Hello with a newline at the end gives a different value from Hello.
• Comparing with a file checksum when the text was typed in a form. A file with Windows line endings has two bytes per break and a typed text may have different breaks, so the byte counts can differ.
• Ignoring letter case. Hello and hello are different bytes and give different checksums.
• Expecting the same number as the CRC or other checksum algorithms. BSD sum-16, sum -s (System V), CRC-32 and MD5 are different algorithms with different results.
• Using it to prove a file was not altered on purpose. It is easy to make two different texts with the same value, so use SHA-256 for integrity against tampering.